In response to increasing competition, financial institutions, like other service providers, have begun to adopt sales and service techniques that have been successful in other fields. Marketing financial services poses unique challenges, however. To begin with, most people do not shop for financial services. Instead, something in a customer's life occurs to cause a customer to make a change or be open to a change. There are moments in life when inertia is overcome; either by moving, death, formation of a family, or when a customer becomes angry enough to make a change. For these reasons large unfocused marketing of financial services is usually not effective. Instead, marketing of financial services must be targeted to those inclined to make a change or open other accounts. In the past, accurately identifying customers that are open to change and predicting when these events will occur has been difficult, if not impossible. Thus, there is a need for a better system and method for predicting when customers or potential customers will be open to change.
To anticipate a customer's needs and support targeted marketing, a service provider must know its customer. Knowing one's customers is also important for improved customer service, another proven way of getting and keeping new customers. Since knowing one's customers becomes more difficult when the number of customers increases and the frequency of each customer's contact with a particular employee decreases, the size of a large financial institution's customer base can present an obstacle to some marketing efforts. In the financial community today, a large financial institution may have several million households and customers each with a unique set of accounts. The data available for these households, customers, and accounts is so massive, that it has heretofore not been fully used for marketing campaigns.
In an effort to deal with a large customer database, businesses traditionally maintain customer records. In some cases these records are in the form of simple paper records, but recently electronic records have become common. Originally, separate data storage was used for each electronic record keeping application. Thus, each department in a financial institution, for example, would have a program that created and maintained records needed for its purpose. The problem with this approach is that information must be extensively duplicated. For example, a customer's name and address might appear in separate files in several separate departments.
There are other problems with application specific data storage. Since a customer's information is entered in more than one file, any change in status must be entered into each file, often by different people. Over time the accuracy and uniformity of the data deteriorates. In addition, the use of application specific data storage requires more data entry and more storage space.
The concept of a database, introduced more than twenty years ago, has come a long way toward eliminating these problems. In a database, data is stored in a central location so that there is no duplication of data. Database management programs are used to manage databases. Examples of currently available database management programs include DB2 (for larger databases) and dBase (for personal computers).
Typically, a database management system (DBMS) is used to manage the creation, storage, access, updating, deletion, and use of a database. A typical DBMS creates databases and their structures; provides the means for the control and administration of the data in the database; provides the means for users and application programs to access, enter, modify, and manipulate the data in a database; provides a report generator; provides "ad hoc" query facilities; provides reports to management on who accessed the database and what activity was performed; provides reports to operators on hardware utilization, status of current users, and other monitoring data; and provides automatic backup and recovery routines for the data in databases.
Multiple-user databases present several additional challenges. These include maintaining system performance as the number of users increases, controlling concurrent access of data, maintaining security, and administrating the database.
There are four basic database models: (1) hierarchical, (2) network, (3) relational, and (4) object-oriented. The hierarchical and network models use files for storing data. Data relationships in the hierarchical databases follow hierarchies, or trees, which reflect either a one-to-one or a one-to-many relationship among the record types. Data relationships in network databases follow a many-to-many relationship among the records. The data relationships must be defined at the time that a hierarchical or network database is created. Relational databases use tables for storing data. The data relationships can be dynamically determined by the users and do not have to be defined when the database is created. A relational database uses a database query language for users to access and manipulate data in the database. Query by example and structured query language are two database query languages. Object-oriented databases store data together with procedures in objects.
A relational database is composed of many tables in which data are stored. Tables in a relational database must have unique rows, and each cell or field must contain only one item of information, such as a name, address or identification number. A relational database management system (RDBMS) allows data to be readily created, maintained, manipulated, and retrieved from a relational database.
In most sophisticated databases, data is retrieved by querying the database. Query languages allow users to locate specific records based on the data that they contain. Known query languages include program-specific languages, structured query languages, natural languages, and query by example. When using query languages, the user typically specifies the rules the program follows to select records to be retrieved. These rules are referred to as criteria. Only those records matching the criteria specified are retrieved.
In a relational database, data relationships do not have to be predefined. Users query a relational database and establish data relationships spontaneously by joining common fields. A database query language acts as an interface between users and a relational database management system.
Two basic query styles are used in a relational database: (1) query by example, and (2) structured query language. In query by example, the database management system displays field information and users enter inquiry conditions in the desired fields.
Structured query language (SQL) is the standard database query language used with relational databases. SQL is part of a DBMS, not a separate stand-alone software program. SQL allows users to create and operate sets of related information that are stored in tables. The core of SQL is its flexibility in querying a database.
This flexibility is possible because of the manner in which data are stored in a relational database. Data are stored in tables that have specific properties. These properties include: (1) one or more named columns, (2) the data in each column are of the same type, (3) zero or more rows (zero rows occur when the table is defined but no data are entered yet), (4) every row is unique, (5) a single data value is contained in the intersection of any column and row, and (6) the order of the columns and rows does not matter.
There are two basic schemes for retrieving data from a database: set orientation and record orientation. Each method has advantages and disadvantages.
A set-oriented database allows the user to focus on the characteristics of the data rather than the physical structure of the data. The user works with data in groups, or sets, or tables, rather than as individual tables. DBMSs that use SQL, such as SQL Server, Oracle, and SQL Base, are set-oriented.
Record-oriented databases access data based on the physical structure of data and indexing. A record pointer permits the user to maneuver through a table one record at a time. It is easy to access successive rows or records in a table. However, the developer of the database management system must write the programming code such that it will loop through every record requested, which is a disadvantage. Examples of DBMSs that use a record-oriented approach are dBASE and Clipper.
In SQL, security is maintained by the granting authority. Authority may be granted to an entire database, certain tables, or certain commands. A database administrator must have access to the entire database so that it can be maintained properly, while a user generally needs access to specific tables or parts of tables. For example, a person might have access to a personnel table but not to the salary column in that table.
Attempts to build and use customer databases have a variety of limitations. In a general sense, these limitations fall into two distinct categories: limitations in the sources and quality of data input into the database and limitations on one's ability to search and retrieve data from the database. In some cases these limitations work in opposition to one another. For example, as one improves the size and quality of a databases, searching and retrieving data from the database becomes more difficult.
In recent years, financial institutions, such as banks, have used targeted marketing (especially direct mailing and telemarketing) to market a wide variety of financial products and services to existing and new customers. To assist these efforts, the banks have used traditional databases containing, for example, customer lists and mailing lists. These traditional targeted marketing sources do not, however, take full advantage of the information available to full service financial institutions.
Full service financial institutions typically offer consumers a wide variety of financial products, including traditional deposit, investment, loan, and mortgage accounts, as well as a variety of financial services, including credit cards, brokerage, direct access, business access, checks as cash, telephone bill payment, and safety check. In addition, financial institutions now typically offer access to financial services through a variety of means, including automatic teller machines (ATMs), customer activated terminals (CATs), screen phones, personal computers configured for banking, personal digital assistants, voice response systems, and smart cards, as well as traditional human bank tellers. Information from these diverse sources provides an unusually complete picture of a customer's financial habits and needs. Thus, the ability to store and retrieve this wealth of data in a meaningful way has enormous commercial potential. Despite this commercial potential, there remains a need for a system and method for assembling a comprehensive database from these diverse sources and retrieving information from the central database in a meaningful and practical way.
There are several deficiencies in currently available systems and methods for assembling customer financial data and retrieving information for use in marketing and customer service systems. To begin with, most users (e.g., bank employees) never learn how to use complex query languages. Mastery of the language requires significant training and skill. Instead, developers write custom applications that are used by users having only a limited understanding of the program. Thus, a user's ability to use a database is often limited by the custom applications written by someone else for their use. Consequently, available large scale database systems typically don't have the flexibility to allow the user, the person most familiar with marketing, to use their own knowledge and experience to select criteria retrieving data from the database for targeted marketing. Instead, users must rely on a set of pre-defined queries that may or may not provide the desired results. As a result, the sales campaigns typically only target easily ascertainable groups of new or existing customers, such as all new customers, or all existing customers with certain types of accounts, etc. Since there has been no effective way to quickly generate and distribute lists of sales leads for very specific groups of people that are most likely to subscribe to new financial services being offered, those customers who most likely need or want the additional products a financial institution has to offer are not always the ones targeted by the sales campaigns. This has resulted in less than satisfactory success rates for marketing campaigns.
In addition, those in charge of marketing are often not given access to a customers' entire relationship with the financial institution or complete demographic information about the customer (i.e., the customer's "profile"). Thus, it is difficult for direct mail and telemarketers to address the targeted customers intelligently, with full knowledge of the customer's background and financial situation. Basic information about existing customers is frequently not available, or the response time required to profile an existing customer is too long. These problems tend to create a poor experience for the customer and less than optimum sales performance.
In addition, the sales performance of bank branches, branch managers, and others in charge of the marketing campaigns has not been analyzed and tracked effectively. A complete indication of sales performance has typically only been available after the sales campaigns are complete and after the results of the campaigns are manually collected and analyzed. This typically required a series of paper-based forms and ad hoc systems that generated relatively slow feedback to sales personnel. Thus, there is also a need for a system to provide up-to-date on-line sales summary reports for each of the products and services marketed by the branches, as well as an indication of performance by the individual sales personnel.
In short, there remains a need for an improved integrated system for identifying sales targets, distributing sales leads, enhancing sales tools, and tracking the performance of large sales campaigns and individual salespersons to maximize customer satisfaction, as well as the profit of the financial institution.